JP4525224B2 - Document management program, document management method, and document management apparatus - Google Patents

Description

  The present invention relates to a document management program, a document management method, and a document management apparatus for searching for a document from a plurality of document groups.

  2. Description of the Related Art Conventionally, an apparatus that registers documents such as documents in a database and performs a full text search using a word included in the document as a keyword is used.

For example, according to the technique disclosed in Patent Document 1, in the case of such a search, a predetermined document is registered in the database in advance in order to reduce the trouble of a user repeatedly performing a search before obtaining an appropriate search result, When a document related to a predetermined search condition is extracted from this database, a word existing in the document obtained by the search using the predetermined search condition is picked up, and a new search word candidate to be added at the next search As above, the picked-up word is presented.
Japanese Patent Laid-Open No. 9-218881

  However, according to the above conventional search technology, for example, the search results obtained according to each search condition are not considered, so even if a search is made by selecting a word picked up as a search word candidate, Convenience is low, such as a situation where the number of documents is “0”.

  The present invention has been made in view of the above circumstances, and an object thereof is to provide a document management program, a document management method, and a document management apparatus that can improve convenience.

The invention according to claim 1 is a document management program, and selects a computer connected to a document holding means for holding a plurality of documents as at least a part of the held documents as a target document group. Target selection means, word extraction means for extracting a predetermined word group from individual documents included in the target document group, and words included in the word group based on the number of documents including each word And selecting means for selecting as search keyword candidates. The selected search keyword candidates are used for a predetermined search process.

Invention of Claim 2 is the document management program of Claim 1, Comprising: In the said selection means, when the number of documents in which each said word is contained exists in the range of a predetermined value, The computer is caused to function so as to select at least a part of words included in the word group as search keyword candidates.

A third aspect of the present invention is the document management program according to the first or second aspect, wherein the selected search keyword candidates are hierarchically grouped based on a correlation value between words included in a word group. As a means, the computer is further caused to function so that words included in the hierarchy or lower are presented to the user as a search condition unit for each hierarchy of the hierarchical word group.

The invention according to claim 4 is the document management program according to claim 3, wherein, in the clustering means, of the search keyword candidates related to the search keyword candidate among the search keyword candidate combinations. When the number of documents obtained as a result of the search satisfies a predetermined document number condition, the computer generates information representing a word group for searching for documents related to the search keyword candidates related to the combination. Is supposed to function.

A fifth aspect of the present invention is the document management program according to the fourth aspect, wherein the ratio of the number of documents obtained as a result of the search to the number of the target documents is determined in advance as the document number condition. It is assumed that the condition is within the range of the values.

A sixth aspect of the present invention is the document management program according to the fifth aspect, wherein the range of the values differs depending on the hierarchical position of the word group.

The invention according to claim 7 is the document management program according to any one of claims 3 to 6, wherein when the word group is presented, a search for including any of the words included in each word group Information regarding the number of documents obtained as a result of the search as a condition is also presented.

Invention of Claim 8 is the document management program as described in any one of Claim 1-7, Comprising: The said document holding means classify | categorizes and hold | maintains a document, The said object selection means is, In response to input of information specifying at least a part of the classification, a document included in the specified classification is selected as a target document.

The invention described in claim 9 is a document management apparatus connected to a document holding means for holding a plurality of documents, wherein the word management means extracts each word group from the held documents, Means for selecting at least a part of the words included in the word group as search keyword candidates based on the number of documents including the words, and the selected search keyword candidates are subjected to a predetermined search process. It is intended to be used.

  Embodiments of the present invention will be described with reference to the drawings. As shown in FIG. 1, a document management apparatus 1 according to an embodiment of the present invention is communicably connected to a document database 2 as document holding means via a network. As shown in FIG. 1, the document management apparatus 1 includes a control unit 11, a storage unit 12, an operation unit 13, a display unit 14, and a communication unit 15.

  The control unit 11 can be realized using a CPU or the like, and operates according to a program stored in the storage unit 12. In the present embodiment, the control unit 11 selects a part of documents stored in the document database 2 as a target document group, and extracts a predetermined word group from each document included in the target document group. . And the correlation value of the words contained in the said word group is calculated from the result of a dependency analysis, for example. Further, the appearance frequency of each word (the number of documents in which each word appears among the documents included in the target document group) is calculated, and at least a part of the words contained in the word group is searched for based on the appearance frequency. Processing to select as a candidate is performed. Further, word groups are recursively grouped based on the appearance frequency and the correlation value, and word groups at each level are set as search keyword candidates. The control unit 11 further provides the selected search keyword candidate to a predetermined search process. Specific contents of the processing of these control units 11 will be described in detail later. The storage unit 12 can be realized by a memory element such as a random access memory (RAM) or a read only memory (ROM), a hard disk, or the like. The storage unit 12 holds a program executed by the control unit 11. The storage unit 12 also operates as a work memory that holds various data used in the process of the control unit 11.

  The operation unit 13 can be realized by a mouse, a keyboard, or the like. The operation unit 13 accepts a user operation and outputs the content of the operation to the control unit 11. The display unit 14 is a display or the like, and displays and outputs information according to an instruction input from the control unit 11. The communication unit 15 is connected to a network, and transmits and outputs data via the network in accordance with an instruction input from the control unit 11. The communication unit 15 receives data that arrives via the network and outputs the data to the control unit 11.

  The document database 2 holds document substance data. The document database 2 of the present embodiment classifies and stores each document using a hierarchical directory structure. For example, when a user newly stores a document, the user stores the document in association with any directory (that is, classified into the directory) included in a predetermined hierarchical directory structure.

  Here, the content of the process of the control part 11 is demonstrated. The control unit 11 according to the present embodiment includes a selection process for selecting a target document group, a word group extraction process for extracting a word group, a correlation value calculation process for calculating a correlation value between words, and the appearance of a combination of words. Appearance frequency calculation processing for calculating the frequency, keyword selection processing for selecting search keyword candidates from word combinations, clustering processing for hierarchically grouping the selected search keyword candidates, and this hierarchical grouping And a presentation process for presenting the search keyword candidates to the user. In addition, the search keyword selection operation presented is accepted, and a search process based on the selected search keyword is executed.

  Hereinafter, the contents of these processes will be described. Upon receiving an operation for specifying a directory to be searched from the user via the operation unit 13, the control unit 11 selects a document below the directory (including a directory below the directory below the directory) as a target document. Select as a group (selection process).

  The control unit 11 extracts words from sentences included in each document included in the target document group. Specifically, while sequentially selecting each document as a processing target, a morphological analysis process is performed on sentences included in the selected document to generate a list of words, and the list is further duplicated from this list. Process to remove the word. Further, the control unit 11 removes words included in a predetermined general word group (so-called stop word group) from the extracted words.

  As an example, for the sentence “Knowledge management tools are continuing to develop along with EIP, SFA, CRM tools, etc.”, “Knowledge management /// tool / has / EIP /// SFA / CRM / tools / etc. / to / to / to / to / development / to / continue / to / where / is /. ”To generate a list of words. When the duplicate word is removed from this, “Knowledge management /// Tool / has / EIP / or / SFA / CRM / etc. / and / to / to / development / to / continue / to / to /. The word “tool” that overlaps is removed. Next, particles included in the general word group (a kind of “no” or “ha”) and words that are supposed to appear in any sentence (for example, a kind of “development”, “continue”) And “knowledge management / tool / EIP / SFA / CRM” is obtained as an extraction result of the word group extraction process.

  Further, for each word included in the obtained word group list, the control unit 11 checks the number of target documents including the word (DF (Document Frequency), so-called appearance frequency), and stores the word appearance frequency information as the storage unit 12. To store. Then, the control unit 11 may remove the word from the list when the ratio of the number of documents (DF) to the number of all target documents exceeds a predetermined value. For example, in the case of the above example, when the number DF / N of the target documents including the word “tool” in the total number N of target documents exceeds 0.5 (50%), for example, The word “tool” may be removed from the word list of each document.

  The control unit 11 stores the list of word groups obtained in this way in the storage unit 12 in association with the information specifying the extraction source document.

  Next, the control unit 11 calculates a correlation value between words for each word included in the list of word groups obtained for each document. Here, the correlation value can be calculated by, for example, dependency analysis processing. Specifically, methods such as skeletal structure analysis (for example, Yasuaki Hyodo, Naoshi Ikeda: Analysis of skeletal structure of Japanese long sentences using surface information and N-neighbor blocking method, Information Processing Society of Japan, Vol. 36, No. 9 , pp2091-2101 (1995)), and since the known method can be used as the correlation value calculation method, detailed description thereof is omitted here. By this processing, as shown in FIG. 2, a correlation value related to an arbitrary word pair included in the word group is calculated.

  Here, the control unit 11 generates word combinations while referring to the correlation values. Specifically, first, words are grouped in descending order of correlation. However, when the DF / N of the grouped words exceeds a predetermined value, the grouping is not performed. Further, when the number of words included in the group exceeds a predetermined value, for example, 10 words, the grouping is not performed. Grouping continues until DF / N or the number of words contained in the group cannot be grouped. Alternatively, a threshold value may be set in advance, and among the combinations of words in the group, only those having the lowest correlation value exceeding the threshold value may be grouped.

The same processing is recursively performed on the word group so that the highest group included in the list becomes a predetermined value, for example, 10 groups or less. When grouping groups into higher groups, the upper limit of DF / N is increased. Similarly, when the lower limit value of the correlation is set, the lower limit value of the correlation is lowered.

  Specifically, if the correlation values for the words A to D are calculated as illustrated in FIG. 3A, for example, the word having a correlation value equal to or greater than the threshold value “0.5” from here. Are extracted and grouped in descending order of correlation value. First, the word B and the word D having the largest correlation value are grouped. Next, the grouping of the word A and the word B will be considered, but the group A is not grouped because the correlation value between the word A and the word D is smaller than the threshold value. Next, the grouping of the word C and the word D is considered. Since the correlation value between the word B and the word C is equal to or greater than a threshold value, the word B, the word C, and the word D are grouped.

  The control unit 11 stores the combinations of words thus obtained in the storage unit 12 as keyword candidates.

  Then, the control unit 11 performs a search for all word groups at all levels according to a search condition indicating that any word included in the group is included, and calculates the number of documents (hits) obtained as a result. Check and do not create a word group in which the number of documents does not satisfy a predetermined document number condition.

Specifically, the document number condition is that the ratio of the hit number (DF) to the number of target documents (N) is within a predetermined value range, that is, MIN <DF / N <MAX.
It is a condition.

  Here, the MIN is provided to prevent the number of documents obtained as a result of the search from becoming “0”. However, if there is a case where there is no result of the search, the MIN is provided. Should be set to “0”.

  The control unit 11 stores a list of words or word groups satisfying the document number condition in the storage unit 12 as keyword candidates.

  When the keyword candidate is a word group, the document number condition may be different depending on the level of the group. For example, in the case of a single word, MAX is set to “0.5”, and MAX is increased every time the group level (hierarchy position) is increased. .75 ”,“ 0.875 ”in the case of further grouping the group including the word directly (second level),“ 0.9375 ”in the case of further grouping the group (third level). Also good.

  In this way, the control unit 11 groups words actually included in the target document and presents them to the user. That is, the control unit 11 displays the word list obtained by the above process on the display unit 14. Specifically, the control unit 11 displays a list of word lists and a search button on the display unit 14 as shown in FIG. When the user selects a word or word group from the list of the word list by a mouse click operation or the like and further clicks the search button, the control unit 11 accepts the operation and searches for the selected word or word group. Search results based on the conditions (such as a list of documents resulting from the search) are presented on the display unit 14.

  In addition, when a keyword candidate is selected, if the keyword candidate is a word group, a list of words or word subgroups constituting the word group may be further presented (FIG. 4B). . FIG. 4B shows a case where the word group composed of A, B, and C is obtained by grouping the word groups of A, B, and A, C. In this case, when the search condition “A, B, C” is selected, the search condition “A, B” and the search condition “A, C” are presented individually. When the user selects the search condition “A, B” presented here and clicks the search button, the control unit 11 searches the search result (search result based on the selected search condition “A, B”). A list of documents to be displayed) is presented on the display unit 14.

  Further, the control unit 11 may also present information regarding the number of documents obtained as a result of the search according to each search condition for each search condition to be presented. For example, FIG. 5 shows an example in which the information is presented in association with a list of search conditions. In the example of FIG. 5, the user adds an arbitrary word using the selected word or the like as a search condition unit (conditions such as logical sum (OR) and logical product (AND) for the selected search condition). Can be combined).

  Further, in the example of FIG. 5, when the “AND search” button is clicked, the control unit 11 presents a plurality of search condition lists and generates a new search condition by a logical product of the plurality of search conditions. I can give instructions. In the example of FIG. 5, when a plurality of search conditions are combined by logical sum to generate a new search condition, a plurality of search conditions combined by the logical sum may be selected from a list of search conditions. As described above, even when the logical sum and logical product of the search conditions are specified, information regarding the number of documents in the search result under the specified conditions may be presented.

  Further, in the example of FIG. 5, an interface for selecting a directory to be searched is also presented (X). Each time a directory to be searched is specified in this interface, the above processing for generating a search condition is performed using the documents included in the specified directory as the target document, and the generated search condition is presented. Will be.

  Furthermore, when the user can input an arbitrary word as in the example shown in FIG. 5, when the user inputs a word, the input word includes the input word. Information about the number of documents may be presented. For example, when the appearance frequency information related to the input word is stored in the storage unit 12, with reference to the information, when the appearance frequency (DF) of the word is less than a predetermined frequency threshold, A word such as “The word has few hits or does not hit” is displayed.

  Further, when it is not less than the frequency threshold, the appearance frequency value (DF value) may be displayed (FIG. 6). Furthermore, when the user inputs a word and the word is included in the presented search condition, the word included in the search condition may be selected instead of the input word (FIG. 7). ). In the example of FIG. 7, the word input in the input field remains displayed. However, after selecting the word included in the search condition, the contents of the input field are deleted and input The column may be empty.

  In addition, as a method for presenting the search result, the control unit 11 presents a directory below the directory selected as a search target (a directory immediately below the selected directory) or a list of documents. A document that satisfies the search condition or a directory that includes a document that satisfies the search condition (including the case where it is contained in a lower directory) may be distinguished from other documents or directories and highlighted. (FIG. 8). FIG. 8 shows an example in which a document satisfying the search condition and a directory containing a document satisfying the search condition are shaded and emphasized. In addition to this, the character color or the background color is changed or blinking is performed. For example, the display mode may be differently highlighted.

  According to the present embodiment, search keyword candidates are presented based on the words actually included in the document group to be searched, and the search results obtained according to the presented search keyword candidates are displayed. Considered. For this reason, it is supported that a significant search operation can be easily performed, and convenience can be improved.

It is a block diagram showing the example of a structure of the document management apparatus concerning embodiment of this invention, and the example of a connection to the document database. It is explanatory drawing showing the example of storage of the calculation result of a correlation value. It is explanatory drawing showing the example which produces | generates the combination of a word from a correlation value. It is explanatory drawing showing the example of presentation of search conditions. It is explanatory drawing showing another example of presentation of search conditions. It is explanatory drawing showing the example of a screen when the word for a search is input arbitrarily. It is explanatory drawing showing another example of a screen when the word for a search is input arbitrarily. It is explanatory drawing showing the example of presentation of a search result.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Document management apparatus, 2 Document database, 11 Control part, 12 Storage part, 13 Operation part, 14 Display part, 15 Communication part.

Claims (5)

A computer connected to a document holding means for holding a plurality of documents,
A target selecting means for selecting at least a part of the held document as a target document group;
Word extraction means for extracting a predetermined word group from individual documents included in the target document group;
Means for examining the number of documents containing the word for each word contained in the word group;
Selection means for selecting at least a part of words included in the word group obtained for each document as a search keyword candidate based on the number of documents examined ;
Means for calculating a correlation value between words included in the word group for the selected search keyword candidates ;
Means for grouping the selected search keyword candidates whose calculated correlation values exceed a predetermined threshold in descending order of the calculated correlation values ;
Means for recursively performing the grouping process on the group obtained by the grouping;
Means for presenting a list of search keyword candidates grouped to the user;
Document management program characterized by functioning as The document management program according to claim 1,
In the selection means, when the number of documents including each word is within a predetermined value range, a computer is selected so that at least a part of the words included in the word group is selected as a search keyword candidate. A document management program characterized by functioning. The document management program according to claim 1 or 2,
In the means for grouping the search keyword candidates, the number of documents obtained as a result of searching for documents related to the search keyword candidates related to the combination of the search keyword candidates is determined in advance. A document management program that causes a computer to function to generate information representing a word group for searching for a document related to a search keyword candidate related to the combination when the document number condition is satisfied. The document management program according to claim 3,
The document number condition is a condition that a ratio of the number of documents obtained as a result of the search to the number of the target documents is in a predetermined value range. A document management apparatus connected so as to be accessible to a document holding means for holding a plurality of documents,
A target selecting means for selecting at least a part of the held document as a target document group;
Word extraction means for extracting a predetermined word group from individual documents included in the target document group;
Means for examining the number of documents containing the word for each word contained in the word group;
Selection means for selecting at least a part of words included in a word group obtained for each document as a search keyword candidate based on the number of documents examined ;
Means for calculating a correlation value between words included in the word group for the selected search keyword candidates ;
Means for grouping the selected search keyword candidates whose calculated correlation values exceed a predetermined threshold in descending order of the calculated correlation values ;
Means for recursively performing the grouping process on the group obtained by the grouping;
Means for presenting a list of search keyword candidates grouped to the user;
A document management apparatus comprising:

Images